MySQL High-Availability and Scale-Out architectures · 2013. 3. 17. · MySQL Cluster •...

1Copyright 2007 Oli Sennhauser The World’s Most Popular Open Source Database

MySQL High-Availability andScale-Out architectures

Oli SennhauserSenior Consultant

[email protected]


Introduction

• Who we are?• What we want?


Table of Contents

• Scale-Up vs. Scale-Out• MySQL Replication• MySQL HA solution• MySQL Cluster• Application architectural solutions• Other architectural stuff• It's your turn...• Let us build a replication


When do we think about architecture?

• Performance problems• HA requirements


Scale-up vs. Scale-Out

• Scale-up

• Scale-out

t

Relaxation of constraints


MySQL Replication

• The MySQL answer: Master-Slave-Replication:

Even Oracle uses the expression “scale-out” now!

Master

Slave1 Slave2 Slave3

Application

LB

...

ro

rtw

2 types ofsessions!!!


MySQL Replication in detail

M S

Appl

bin-logbin-log... relay-log

IO thread

SQL thread

asynchronous!

relay-log

rtwro

• wide distance possible!• thin line possible!


MySQL Replication varieties

M

SI SI SI

• Cascaded replication:

for example Yahoo!

S1 S2 S3 S4 S6S5 S7 S8

...

...

black hole SE!


MySQL Replication varieties

M

S2 S3

• HA with “hot standby” Slave:

• active – active fail over!

S1

Sbackup Sstandby

Appl

VIP HA solution


MySQL HA solution

M

S2 S3

• HA with MySQL:

• 99.99% HA (four nine)• active – passive fail over!

S1

M'

Appl

VIP

DRBD

heartbeat


MySQL HA solution

M

S2 S3

• HA with MySQL and SAN:

S1

M'

Appl

VIP

SAN

heartbeat


Multi-Master Replication

M1

S12 S13

• actually Master-Master Replication

• This architecture does NOT solve your write problems!!!

S11

Appl

M2

Appl

VIP/LB

S22 S23S21


Multi-Master Replication

M1

S12 S13

• HA Multi-Master Replication

S11

M2

S22 S23S21

M2'

Appl

VIP

DRBD

M1'

Appl

VIP

DRBD

VIP/LB


Shared disk cluster

DB1

• This is in 99.9% of the cases NOT your solution!!!• It looks like Oracle RAC but MySQL does NOT

(yet) have the instruments needed!• Can be useful in some special kinds of reporting.

DB2

shareddisk

DB2ro!


Where are we now?

• The architectures above solved:– your read problems.– some HA requirements.

• But they did not solve:– your write problems!

• And what when we need:– higher HA?– synchronous replication?– more write speed?


MySQL Cluster

• Shared-nothing architecture• Synchronous replication (2-Phase commit)• Fast automatic fail over• High performance (also writing)• High transactional throughput• No special components required• In-Memory database (in 5.1 also disk support)• Scalable, 1000's of transactions per second• 99.999% HA (five nine)• On-line upgrade path (at least on GA within same

version)


MySQL Cluster architecture

Application Application Application

MySQLServer

NDB Cluster(Data nodes)

MySQLServer

MySQLServer

DB DB

DBDB

MGM Server

MGM client

ApplicationApplicationApplication

config.ini

my.cnfSQLNodes

DataNodes

MgmtNodes


MySQL Cluster HA features

MySQLServerMySQLServer

DataNodeDataNode

DataNodeDataNode

NDBStorage Engine

ManagementServer

ManagementServer

NDB APINDB API

MySQL Cluster

DataNodeDataNode

DataNodeDataNode



ManagementServer

ManagementServer



X X X X

X X X


Cluster with Replication

S2 S3

• for read scale-out (Reporting):

S1

Appl

ndbd ndbd

ndbd ndbd

mysqld mysqld mysqld

Reporting

rw



• Cluster-Cluster replication for disaster fail over (MySQL 5.0):

Appl

ndbd ndbd

ndbd ndbd


rw

ndbd ndbd

ndbd ndbd


fail over



• Cluster-Cluster replication for disaster fail over (MySQL 5.1):

Appl

ndbd ndbd

ndbd ndbd


rw

ndbd ndbd

ndbd ndbd


fail over



• circular Cluster-Cluster replication (>= MySQL 5.1.18?):


Cluster examples

• Session handling• Telecom (Mobile)• VoIP• RSS-Feed aggregation• Mail• On-line Games

• Use cluster where– you need HA– you have high write load– you do little Joins and Grouping


How to go on?

• Read is a caching problem!➔ More RAM or scale-out.

● Write is a batching problem!➔ Batch your load,➔ Buy stronger I/O system.➔ Use MySQL Cluster.➔ SSD!

● What then?


SSD disruption

• I/O system without any movable parts!• SSD = Solid State disk (Flash memory, NAND,

NOR chips, RAM-SAN).• During the last year(s)• Price from 15'000 USD / 160 Gbyte -> 100 USD /

32 Gbyte• 10-50 times faster than mechanical I/O systems• 1 Mio write cycles dead :-( (special FS!)

• RAM: Huge amount of memory is cheap!

➔ This will disrupt the database world!


What then?

• With or without SSD we will reach a physical limit. What then?

• Application partitioning:– Split applications

• OLTP vs. OLAP• all in one

– Segment your application


Application partitioning

• OLTP vs. OLAP– hot business data (trx) vs. “old” reporting data

• “All in one”– Sessions, user tracking, ads, chat, booking

• Segment– split by for example 1 Mio users (split by user_id).


Architecture examples


Some more architecture stuff?

• Databases are slow!!!➔ Whenever possible try to avoid databases! Use

memcached for example.• The SP trap.

➔ Stored Procedures are a lock in! Try to avoid SP!• Use Materialized Views (MV) and/or shadow

tables.• VM/SAN is nice for consolidation but not for

performance/scale-out!• Backup and Staging• MySQL – Proxy


Now it's your turn...

• Your problems?

• Let us build a replication...

Date post:	02-Feb-2021
Category:	Documents
Upload:	others
View:	4 times
Download:	0 times

MySQL High-Availability and Scale-Out architectures · 2013. 3. 17. · MySQL Cluster •...

Documents